Optimal weighting scheme for affected sib-pair analysis of sibship data

Area of Science:

• Genetics
• Statistical genetics
• Bioinformatics

Background:

• Affected sib-pair (ASP) analysis is crucial for genetic linkage detection.
• Weighting schemes are necessary for sibships with varying numbers of affected/unaffected members.
• Existing methods use independent pairs or Shannon information for weighting.

Purpose of the Study:

• To determine optimal weighting schemes for ASP linkage analysis.
• To maximize the power of detecting genetic linkage.
• To compare existing weighting schemes against theoretically optimal ones.

Main Methods:

• Derived theoretical means and variances of sibship contributions under null and alternative hypotheses.
• Utilized a generalized single locus model.
• Calculated optimal weights based on derived theoretical values.

Main Results:

• Identified weighting schemes that maximize linkage detection power.
• Found that equal weighting of sib-pairs is near-optimal in most cases.
• Determined that sibships with five or more affected siblings require adjusted weighting due to substantial parental homozygosity probability.

Conclusions:

• Optimal weighting schemes can be derived from theoretical means and variances.
• Near-optimal power is achieved with equal weighting for most sib-pairs.
• Sibship informativeness in ASP analysis is more related to the number of affected sib-pairs than affected siblings, up to five.